How many matrices agree with the mean?

Hello,

So I have 11 matrices of 64x128x12 named A,B.....etc, with the first eleven being one variable and the second eleven being another.

After calculating the mean of these matrices (again 64x128x12), what would be the best way/tools to calculate how many of the eleven matrices agree with the mean (within 95% confidence)?

I would like my output to be something like 8 out of 11 matrices agree in these 'locations'.

Thank you!

Answer 1

[Star Strider]

The best estimate of the dispersion of the mean is to use the standard error of the mean:

where σ is the standard deviation and N are the number of observations. You can then multiply that by the critical values for the normal distribution [-1.96 1.96] to get the 95% confidence limits. This would work if you are taking the mean across the 64 or 128 dimensions, if you are using the 12 dimension, you would need to use the tinv function to get the critical values from the t-distribution.

However, what your Question lacks in clarity it makes up for in confusion. Please provide more detail as to what you are doing, and what the dimensions of the result of your taking the mean are.